The MCF-10A and MDA-MB-231 cell lines were purchased from the American Type Culture Collection (ATCC) and cultured according to the manufacturer’s directions. MCF-10AT was obtained from the Barbara Ann Karmanos Cancer Institute. MCF-10A was derived from spontaneous immortalized breast epithelial cells of a patient with fibrocystic disease. MCF-10AT was obtained by H-ras-transfected MCF-10A cells. MDA-MB-231 was an estrogen receptor-negative breast cancer cell line. MCF-10A and MCF-10AT were maintained in DMEM/F12 medium supplemented with 10% horse serum, 20 ng/ml EGF, 10 μg/ml insulin, and 50 μg/ml hydrocortisone. MDA-MB-231 was cultured with Leibovitz’s L-15 medium containing 10% FBS. All cells were incubated in the 37°C incubator with 5% CO₂ and 95% humidity.

Cells were seeded into six-well plates at the concentration of 500 cells per well and maintained with or without β-BA for 2 weeks. For analysis, the plates were washed with PBS twice, fixed with methanol for 30 min, stained by 0.1% crystal violet, and then counted via ImageJ.

The effect of β-BA on cells was assessed by CCK8 assay. Twenty-four hours before treatment, 4 × 10³ cells were seeded into each well, followed by adding various concentrations of β-BA. On the day of detection, the supernatant was replaced by 10% CCK8 mixture (10 μl of CCK8 with 90 μl of basic medium for each well). With additional 2 h incubation at 37°C, the plate was read at 450 nm by a microplate spectrophotometer.

The apoptosis assay was determined using an Annexin V-APC/PI detection kit (DOJINDO, Japanese) following the manufacturer’s protocol. The cells under detection were harvested, washed, and suspended in the binding buffer. After staining with the Annexin V/PI staining mixture for 15 min at room temperature, the samples were analyzed by flow cytometry (Beckman, USA) within 1 h.

To identify potential targets and signaling pathways, β-BA (PubChem CID:168928) was submitted to Bioinformatics Analysis Tool for Molecular mechanism of TCM (BATMAN-TCM, http://bionet.ncpsb.org/batman-tcm/). The predicted targets were submitted for Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis on the DAVID website to enrich related signaling pathways.

Extracellular acidification rate (ECAR) was analyzed using Glycolysis Stress Test Kit via Seahorse XF96 instrument (Agilent Technologies, USA). Briefly, cells were pre-treated with or without β-BA for 48 h and then seeded into a plate one day before testing. During the experiment, cells were treated with glucose, oligomycin, and 2-deoxy-glucose sequentially with ECAR value quantified at different time points.

2-NBDG (Med Chem Express, USA), a fluorescent D-glucose analog, was used to detect glucose uptake level of cells following treatment with β-BA. The stock solution was diluted with basic medium to yield 20 μM working solution in advance. After treatment with 40 μM β-BA for 48 h, 2-NBDG working solution was added to cover the surface of culture cells sufficiently and incubated at 37°C for 30 min. Then, the cells were washed with PBS twice to remove residual reagent and detected as soon as possible. The intensity of fluorescence was detected using flow cytometry (Beckman, USA) with FL1.

For cells, reasonable RIPA lysis buffer was added to the dishes following twice washing with PBS. Then, cells were harvested by gentle scraping and stored at −80°C if not processed immediately. The well-prepared total protein was separated by 12% or 10% SDS-PAGE and transferred to a 0.22-μm polyvinylidene fluoride (PVDF) membrane (Millipore, Billerica, USA). After blocking with 5% skimmed milk at room temperature, the membrane was washed 3 times and incubated with a certain first antibody overnight at 4°C. HRP-conjugated secondary antibody (Cell Signaling Technology, Inc., Boston, USA) was used with bands visualized by the ECL detection kit.

RNA was extracted from control and drug-treated cells using Trizol reagent (Invitrogen, Carlsbad, CA, USA) and reverse transcribed (Takara, Japan). Then, the cDNA was submitted to quantitative PCR using 2× SYBR green master mix (Bio-Rad laboratories, USA) on a CFX96 detection system according to the manufacturer’s instructions. Data of samples were normalized by the levels of ACTB. The sequences of primers were acquired from PrimerBank.

The CDS sequence of GLUT1 was successfully cloned into pcDNA3.1 plasmid and validated by sequencing. Then, the plasmid pcDNA3.1-GLUT1 and its control were transfected into MCF-10AT using Lipofectamine 2000 (Invitrogen, CA, USA) according to the protocol of the manufacturer. Twenty-four hours after transfection, the cells were treated with vehicle or β-BA 48 h before downstream assays.

DMBA was precisely weighted, dissolved in sesame oil, and ultrasonic processed in a water bath at a constant temperature of 60°C to get a homogeneous mixture. The concentration of DMBA was 7 mg/ml.

The blank matrix was heated to melt and slowly mixed with Tam via constant stirring. Then, the mix was cooled down to room temperature. The concentration of Tam in cream was 5 mg/g.

The blank matrix was prepared as previously described. Briefly, the water phase was prepared by dissolving 7 g of KOH in an appropriate amount of distilled water and heated to 80°C. Stearic acid (140 g) was heated, melted, and cooled down to 80°C. KOH water solution (100 g) and glycerol were added and stirred constantly. Then, an appropriate amount of distilled water was added, and the volume was metered to 1,000 ml to obtain the blank matrix (Mat). The matrix was continuously stirred and cooled down to 40°C. Then, a definite dose of β-BA (MedChemExpress, USA) was added, thoroughly mixed, and cooled down to room temperature to obtain the β-BA cream. The concentration of β-BA in cream was 5 mg/g.

Forty female SD rats that were 6 weeks old were fed in the SPF-grade fostering environment of Jinan University Animal Center. Then, the rats were randomly divided into four groups (10 per group): the vehicle group, the disease model group, the Tam-treated group, and the β-BA acid-treated group. One-week post arrival, the hair around the breast (1 cm diameter) of these rats was removed by depilatory paste for follow-up treatment. Finally, the rats were sacrificed and tested in the 14th week. This animal experiment was approved by the Laboratory Animal Ethics Committee of Jinan University. The experimental procedure and animal welfare strictly followed the Care and Use of Laboratory Animals (Ministry of Science and Technology of China, 2006) and related regulations of Jinan University.

Once separated from the bodies of rats, organ tissues were fixed in formalin immediately, and later embedded in paraffin. The blocks were cut into 5-μm sections. Before staining, the slides were baked, dewaxed, and hydrated. For HE staining, the slides were stained with hematoxylin first for 2 min followed by hydrochloric acid alcohol differentiation and then eosin stain.

Error bars represent the S.D., as indicated in figure legends. All statistical tests were analyzed by Student’s two-tailed t-test for comparison of two groups and by analysis of variance (with post-hoc comparisons using Dunnett’s test) for comparison of multiple groups using GraphPad Prism v. 8.0 (La Jolla, CA). p < 0.05 was considered statistically significant.

The chemical structure of cordycepin is shown in Figure 1A. Cell lines used in this experiment involve MCF-10A, MCF-10AT, and MDA-MB-231. Due to the various similarities between precancerous lesions and tumors, a commonly used triple-negative breast cancer cell line, MDA-MB-231, was used as a control. These cells were treated with β-BA at various concentrations for 24, 48, or 72 h to investigate the toxic effect of β-BA on these cells. As shown in Figure 1B, the CCK-8 assay showed that β-BA displayed a significant inhibition effect on MCF-10AT and MDA-MB-231, and a mild effect on MCF-10A. The IC₅₀ values of β-BA at 48 and 72 h were 89.04 and 91.41 μM in MCF-10A cells, 37.2 and 30.3 μM in MCF-10AT cells, and 39.4 and 26.23 μM in MDA-MB-231 cells, respectively. In addition, a colony formation assay was performed to assess the inhibitory effect of β-BA. The results revealed that MCF-10AT and MDA-MB-231 treated with 5, 10, and 20 μM β-BA for 2 weeks formed fewer and smaller colonies than control (Figure 1C), indicating that β-BA inhibited the growth of these cells in a dose-dependent manner, while MCF-10A was suppressed in a relatively milder way.

We further explored whether β-BA had an effect on the apoptosis of MCF-10A and MCF-10AT. After treating with β-BA for 48 h, MCF-10A and MCF-10AT cells were analyzed via Annexin V/PI staining. As shown in Figures 2A, B, the Q3 area indicating early apoptosis and the Q2 area indicating late apoptosis were both dramatically increased for MCF-10AT, suggesting that β-BA induced significant apoptosis, while β-BA induced apoptosis on MCF-10A in a relative milder way. To further confirm these results, we performed Western blot to examine protein expression of apoptosis-related genes. Activation of cysteine-aspartic acid proteases (caspases) plays an important role in cell apoptosis. As shown in Figure 2C, the cleaved form of caspase 3 was dominantly increased in MCF-10AT. Bax, the core regulator of the intrinsic pathway of apoptosis, also increased following β-BA treatment in MCF-10AT. Altogether, β-BA markedly promoted apoptosis in MCF-10AT with a relatively milder effect on MCF-10A.

To further explore the potential mechanism of β-BA in inhibiting proliferation and inducing apoptosis on MCF-10AT, we searched the BATMAN-TCM website to predict potential targets. A total of 45 genes were obtained, including COX1, ADH1B, and CYP17A1. To further characterize the potential affected pathways, these genes were submitted for KEGG analysis. The top enriched pathway was the metabolic pathway (Figure 3A). Considering glycolysis increase is a key character of precancerous lesions compared to normal tissues, and we detected glycolysis changes of MCF-10A and MCF-10AT via seahorse assays. Following treatment with β-BA, MCF-10AT displayed decreased extracellular acidification rate (ECAR) compared to control cells, while the change of ECAR in MCF-10A was mild (Figure 3B). Consistently, decreased lactate production was observed in β-BA-treated MCF-10AT (Figure 3C), and the ATP level was also decreased (Figure 3D). Since ATP is the main source of energy required by most in vivo biochemical processes, its reduction may limit cell survival. The AMPK system is a key player in regulating energy balance, which could be activated by increased AMP/ATP ratio. Once activated, it switches anabolic processes to catabolic pathways. Here, we found that the phosphorylation of AMPK α-subunit was enhanced following β-BA treatment with the total expression of AMPK not influenced. The activation of AMPK downstream target ACC further confirmed these. mTORC1 as a reported target of AMPK functions to promote cell growth and proliferation. Here, the results of Western blot showed that p-mTOR as well as the downstream target p-p70S6k of MCF-10AT were reduced following β-BA treatment, with the pan-protein expression not affected (Figure 3E). Altogether, β-BA could suppress the glycolysis process and induce lactate and ATP reduction in MCF-10AT. The energy stresses would activate the AMPK pathway and inhibit the mTORC1 pathway, which may partially explain the inhibition effects of β-BA on MCF-10AT.

To obtain further insight into the mechanisms of β-BA in regulating the glycolysis process, we used docking analysis to search potential targets of β-BA. To focus on the glycolysis process, we predicted the binding between β-BA and the key proteins in the glycolysis pathway. The molecular docking analysis was based on vital parameters like binding site and strength of molecular interactions. The initial structures were taken from the Protein Data Bank (PDB). Substrates were docked into the active site of protein using the AutoDock Vina tool (15) in Chimera (16). The docked poses with the highest docking scores were used for the subsequent studies. As a result, the highest affinity was reached between β-BA and GLUT1 (−10.5 kcal/mol), which was higher than the redocking score between natural substrate and GLUT1 (−6.9 kcal/mol) (Table 1). GLUT1 is a key transporter that regulates glucose absorbance and maintains intracellular glucose level, which governs the fuel of glycolysis. It was reported that Phe379, Trp388, and Trp412 played essential roles in GLUT1 function and glucose uptake (17–19). Notably, our results showed that β-BA bonded to the same pocket and interacted with these reported residues as natural substrate glucose (Figures 4A, B), which indicated that it might be a competent inhibitor for GLUT1. To validate the effect of β-BA on GLUT1, we checked the expression of GLUT1. As a result, neither the protein nor the mRNA level of GLUT1 in MCF-10AT was influenced following β-BA treatment (Figures 4C, D), while the results of 2-NBDG glucose uptake assay indicated that β-BA suppressed glucose transport (Figure 4E). In summary, β-BA could bind with GLUT1 to block glucose uptake, without influencing GLUT1 expression in MCF-10AT.

GLUT1 was reported to be upregulated in various types of carcinomas, while its function in breast precancerous cells remains to be elusive. We overexpressed GLUT1 by transfecting plasmid pcDNA3.1-GLUT1 into MCF-10AT and then tested the efficiency (Figure 5A). As shown in Figures 5B, C, forced expression of GLUT1 promoted cell proliferation and increased glucose uptake.

We further checked whether β-BA-regulated glycolysis alteration was related to GLUT1. The results suggested that overexpression of GLUT1 could predominantly reverse the proliferation inhibition induced by β-BA (Figure 5G). Consistently, overexpression of GLUT1 can also rescue the glycolysis phenotypes in MCF-10AT cells (Figures 5D–F). These results confirmed that GLUT1 was involved in β-BA-regulated glycolysis of MCF-10AT.

To further evaluate the therapy value of β-BA in breast precancerous prevention, β-BA was given to a precancerous rat model in the form of cream once a day. The animal model was established by DMBA combined with estrogen and progestin induction in SD female rats, with Tam used as a positive control. After 14 weeks of treatment, all the rats were sacrificed, and the morphology of mammary glands was checked by HE staining (Figure 6). As shown in Table 2, the rats in the normal group hardly showed abnormal hyperplasia, while in the model group, most animals developed atypical hyperplasia and invasive carcinoma (77.5%). After treating with β-BA cream, the dysplasia and carcinoma phenotypes were significantly prevented (44.2%), the effect of which was comparable to the Tam-positive treatment group (36.6%). These results further validated that β-BA could relieve precancerous breast lesion.